1. Field of the Invention
The present invention relates to methods for producing biotinylated proteins in vitro and in recombinant host cells. The invention therefore relates to the field of molecular biology, but given the diverse uses for recombinant proteins, the invention also relates to the fields of chemistry, pharmacology, biotechnology, and medical diagnostics.
2. Description of the Background Art
The ability to synthesize DNA chemically has made possible the construction of peptides and proteins not otherwise found in nature and useful in a wide variety of methods that would otherwise be very difficult or impossible to perform. One illustrative example of this technology relates to the class of molecules known as receptors. Receptor proteins mediate important biological functions through interactions with ligands. For many years, researchers have attempted to isolate and identify ligands that interact with receptors in ways that can help ameliorate human (and other) disease. The advent of molecular biology has revolutionized the way these researchers study receptor-ligand interaction. For instance, standard molecular biology techniques have enabled the cloning and high-level expression of many receptors in recombinant host cells.
The patent literature, for instance, is replete with publications describing the recombinant expression of receptor proteins. See, e.g., PCT Patent Pub. No. 91/18982 and U.S. Pat. Nos. 5,081,228 and 4,968,607, which describe recombinant DNA molecules encoding the IL-1 receptor; U.S. Pat. Nos. 4,816,565; 4,578,335; and 4,845,198, which describe recombinant DNA and proteins relating to the IL-2 receptor; PCT Patent Pub. No. 91/08214, which describes EGF receptor gene related nucleic acids; PCT Patent Pub. No. 91/16431 and U.S. Pat. No. 4,897,264, which describe the interferon gamma receptor and related proteins and nucleic acids; European Pat. Office (EPO) Pub. No. 377,489, which describes the C5a receptor protein; PCT Patent Pub. No. 90/08822, which describes the EPO receptor and related nucleic acids; and PCT Patent Pub. No. 92/01715, which describes MHC receptors.
Several of the above publications not only describe how to isolate a particular receptor protein (or the gene encoding the protein) but also describe variants of the receptor that may be useful in ways the natural or native receptor is not. For instance, PCT Patent Pub. No. 91/16431 describes soluble versions of the gamma interferon receptor, while PCT Patent Pub. No. 92/01715 describes how to produce soluble cell-surface dimeric proteins. This latter technology involves expression of the receptor with a signal for lipid attachment; once the lipid is attached to the receptor, the receptor becomes anchored in the cell membrane, where the dimeric form of the receptor is assembled. See also U.S. Ser. No. 947,339, filed on Sep. 18, 1992, and incorporated herein by reference for all purposes, which describes how HPAP-containing receptors can be cleaved from the cell surface and how the anchoring sequences that remain can serve as recognition sequences for antibodies that are used to immobilize the receptor.
The advances made with respect to receptor cloning and expression have been accompanied by advances in technology relating to methods for screening a receptor against compounds that may interact with the receptor in a desired fashion. One such advance relates to the generation of large numbers of compounds, or potential ligands, in a variety of random and semi-random "peptide diversity" generation systems. These systems include the "peptides on plasmids" system described in U.S. Pat. No. 5,338,665, which is a continuation-in-part of U.S. Pat. No. 5,270,170; the "peptides on phage" system described in U.S. patent application Ser. No. 718,577, filed Jun. 20, 1991, which is a continuation-in-part of Ser. No. 541,108, filed Jun. 20, 1990; Cwirla et al., August 1990, Proc. Natl. Acad. Sci. USA 87: 6378-6382; Barrett et al., 1992, Analyt. Biochem. 204: 357-364; and PCT Patent Pub. Nos. 91/18980 and 91/19818; the phage-based antibody display systems described in U.S. patent application Ser. No. 517,659, filed May 11, 1990, and PCT Patent Pub. No. 91/17271; the bead-based systems for generating and screening nucleic acid ligands described in PCT Pub. Nos. 91/19813, 92/05258, and 92/14843; the bead-based system described in U.S. patent application Ser. No. 946,239, filed Sep. 16, 1992, which is a continuation-in-part of Ser. No. 762,522, filed Sep. 18, 1991; and the "very large scaled immobilized polymer synthesis" system described in U.S. Pat. No. 5,143,854; PCT Patent Pub. Nos. 90/15070 and 92/10092; U.S. patent application Ser. No. 624,120, filed Dec. 6, 1990; Fodor et al., Feb. 15, 1991, Science 251: 767-773; Dower and Fodor, 1991, Ann. Rep. Med. Chem. 26:271-180; and U.S. patent application Ser. No. 805,727, filed Dec. 6, 1991. Each of the above references is incorporated herein by reference for all purposes.
Other developments relate to how the receptor is used in such screening methods. One important advance relates to the development of reagents and methods for immobilizing one or more receptors in a spatially defined array, as described in PCT Patent Pub. No. 91/07087. In one embodiment of this method, a receptor is attached to avidin and then immobilized on a surface that bears biotin groups. The surface is first prepared, however, with caged biotin groups, which will not bind avidin until the caging group is removed by, in this embodiment, irradiation. Once the avidinylated receptor is bound to the biotin groups on the surface, the surface can be used in screening compounds against the receptor.
Biotin is a coenzyme that is covalently attached to several enzymes involved in the transfer of activated carboxyl groups. As the above example illustrates, biotin labeling of molecules not normally biotinylated can be used to label, detect, purify, and/or immobilize such molecules. These methods also rely upon the proteins avidin and streptavidin, which bind very tightly and specifically to biotin and other biotin-binding molecules, some of which bind to biotin with different affinity than avidin. Typically, the biotinylated molecules used in such methods are prepared by an in vitro biotinylation process. A method for biotinylating proteins synthesized by recombinant DNA techniques in vivo would eliminate the need to biotinylate these proteins chemically after purification and would greatly simplify the purification process, due to the ability to use the biotin as an affinity tag (see Green, 1975, Adv. Protein Res. 29:85-133, incorporated herein by reference).
Biotin is added to proteins in vivo through the formation of an amide bond between the biotin carboxyl group and the epsilon-amino group of specific lysine residues in a reaction that requires ATP. In normal E. coli, only one protein is biotinylated, the biotin carboxyl carrier protein (BCCP) subunit of acetyl-CoA carboxylase. This reaction is catalyzed by the biotin-protein ligase (BirA), the product of the birA gene (see Cronan, 1989, Cell 58: 427-429, incorporated herein by reference).
Others have proposed a means by which biotin labeling can be accomplished in vivo by the addition of a domain of at least 75 amino acids to recombinant proteins (see Cronan, 1990, J. Biol. Chem. 265: 10327-10333, incorporated herein by reference). See also Cress et al., 1993, Promega Notes 42: 2-7. Addition of this 75 amino acid domain to several different proteins leads to the biotinylation of the fusion proteins by BirA on a specific lysine of the added domain. Addition of smaller fragments of the 75 residue domain does not lead to biotinylation, implying that a reasonably complex recognition domain is required. Changes in the sequence of biotinylated proteins as far as 33 residues from the modified lysine abolish biotinylation (see Murtif and Samols, 1987, J. Biol. Chem. 262: 11813-11816). Changes close to the lysine also affect biotinylation (see Shenoy et al., 1988, FASEB J. 2: 2505-2511, and Shenoy et al., 1992, J. Biol. Chem. 267: 18407-18412). Unfortunately, however, the addition of such a large protein domain may negatively affect the biochemical properties of a biotinylated protein. Smaller domains that specify biotinylation would be very beneficial, in that such domains would have a minimal structural effect on the wide variety of possible fusion partners. Also, the 75 residue domain does not lead to complete biotinylation of the domain, and improved domains could be more efficient. The present invention provides such improved biotinylation domains.